Once the diyBMS-v4 LEAF is functioning I will give Stuart the files and he can add them to his repository. I made a few changes to his design but nothing serious. This is the first ‘kick at the can’ and I expect I made mistakes with Kicad. It is miserable doing a design with a new pcb cad package and I needed mentally archive commands from 4 other PCB software packages (30 years of misery).to keep from yelling at the monitor… “Do what I want not what I asked you to do”
There is some hand assembly required, JST connectors, ATtiny and power resistors. As soon as the last parts arrive and I can talk to a board I will let the group know of the status. I ordered 20 boards but only need 14 any functioning boards beyond my needs I will offer to the group to play with for the cost of shipping.
George
Ooooh my English is bad, after a few months in other projects, I will start welding the first V4, there is some welding guide, a template to see the values of the resistances I can not extract the gerbera file an optimal size and an adequate font, thanks for the work you do.
Ooooh mi ingles es pésimo, despues de unos meses en otros proyectos, me voy a poner a soldar los primeros V4 , hay alguna guia de soldadura, una plantilla para ver los valores de las resistencias no logro extraer el archivo de los gerbera un tamaño optimo y una fuente de letra adecuada, gracias por el trabajo que haceis.
For anyone who is about to buy some boards or components, the HMHA2801 can be swapped out for “EL3H7(B)(TA)-V” - its physically the same size and pin out.
This part is $0.0948 compared to $0.2580 for the HMHA2801
GeorgeBourdreau thank you, looking forward to seeing the progress and feedback.
Atanisoft, Thank you for the feedback, you make some great points. I have ordered a few boards and plan to look closely at the traces, current carrying capability and most importantly the end-to-end system. All systems are going to need some sort of per cell fuze, especially as you stack more and more parallel cells and want to draw/push more current. I think the Jehu system is pretty slick and even if I have to redesign or modify the PWB, I am going to build a test system to try out. Once I have a complete system with measurements and analysis I can report back for constructive criticism. Per Stuart’s question, I think we can probably all agree that systems (regardless of BMS type) need to be analyzed (and tested) end-to-end to ensure that traces are not too small or BMS wires are not too small for the systems application (current, voltage, #cells, etc…).
I’ve been fighting this myself, even some parts listed as “basic” are showing up as “extended” when sending a BOM/CPL… Virtually all parts smaller than 1206 or 0603 are “extended” regardless of what it shows. It might be a bug but I doubt it.
That isn’t too bad really, with the ATTiny841 via DigiKey it comes out to approximately $3 USD/board.
great influxdb/grafana job! Could you please explain a litte bit more how to get up running influxdb/grafana with diybms v4? Maybe you could write a short guide for beginners (inculding me)?
Particulary I don’t know which parts i’ll need for the integration? A raspberry Pi? And i’m struggling to create the influxDB, because i’m overwhelmed by the settings on the influxdata-website…
I would be happy, if you can give me/us support with the first steps! Thanks!
Hi Stuart
Firstly, nice body of work.
I have a parallel problem to a battery stack - it’s a supercapacitor stack. There are some similarities (need for isolation, voltage balancing, protection from reverse voltage, current limiting), and differences in energy available, the potential for a capacitor to go to zero charge, higher potential current flows.
I’m wondering whether your circuitry and software can be applied.
4 x 16.2V supercap stacks of around 50F;
Stacks will be precharged to 10V, will operate between 10V and 14.5V
Stack may need to be trickled to full 16.2V on occasion if balance is lost.
Precharge is from a diode-protected 300mA supply.
There are six cells per stack.
Cells are protected by MOSFETs with dissipation resistors.
There will be Schottky diodes to limit negative excursions for each cell, and high-amperage Blue LEDs as a second layer of over-voltage power dissipation.
The four stacks comprise an interworking device between an islanding charger and a Redflow flow battery at a nominal 48V, with usual operation between 48V and 57V.
I envisage 4 monitor boards, one per stack.
Can the monitor boards run happily if fed with 10-16V? A simple linear PSU is possible.
What current draw will the monitor boards take per stack?
Can the stack voltage be set between 14V and 16.2V?
Hi, unfortunately the boards won’t operate at that level of voltage. The CPU (attiny) operates at the voltage of the pack and this is a max of 5v, although the board is designed for max of 4.5v
Is there an updated BOM list available to order some partially assembled boards from JLCPCB? I would love to order some boards, but it would be a bit tedious to solder on SMD components for a bunch of boards.
Hi Stuart
Thanks for your fast response. I’ll keep an eye on what you are doing, because your USP is the hardware software integration; there are a few hardware-oriented cell balancers of varying sizes including a couple of power transfer designs based on relatively new chips - however, it’s difficult to know what they are up to :-). Your board linking approach is elegant, and centralising the information means that hooks to existing BMS are a great deal easier, whether for reporting or alarms.
The concept of one board per cell is solid for a battery environment; the other extreme of trailing wires from every cell to a central 12S/24S or bigger board gets potentially very dangerous if there is a wiring overheat issue - a small fire in the sensor wire cluster and everything goes to hell very rapidly.
What I might do downstream is look to build a module sensor board reflecting your design and controller techniques - a microcontroller that can sample six cells and stack, probably having an isolated external power option (reflecting the lesser power and wider voltage range in supercapacitors).
The jlcpcb_assembly branch on GitHub has everything you should need EXCEPT for the ATTiny841 which would still need to be hand soldered (which is surprisingly not that bad with a hot iron and flux).
I would defer to Stuart though before ordering as he has made a couple changes on the branch since I ordered mine. I would also suggest ordering more boards than you need so you have a couple backup boards just in case you need them, I plan on using 14 boards but ordered 30. At that quantity the cost per board is very cheap.
I was planning on ordering a large batch of boards. I need at 32 for my current project. I don’t mind soldering a few components, I assembled some SMD electronic kits before. My biggest fear is ordering a bunch of boards and having to fix an error in each one because I selected the incorrect part.
JLC will only allow up to 30 in an order unfortunately. So you would need to submit two orders (30 and 5). You will end up paying a bit extra due to the “extended” component fees. The other option is order the SMT stencil and assemble yourself which would likely end up cheaper since you would save the “extended” part fees. Time to assemble a PCB yourself would be ~10min per board if you place all components and hot air them. If you opt for JLC assembly you can count on 2-3min per board to solder the THT (four or five parts) and the ATTiny841.
That was my fear as well but Stuart did an amazing job on the jlcpcb_assembly branch when I ordered mine. So far they have all tested fine and quality appears to be very good. I’ll be able to do more testing tomorrow after the JST-PH crimp connectors arrive (forgot to order them with the rest of the parts, oops)